Data Recovery and Disk Utilities Forum @R-TT

Hello,

I am trying to re-create a Linux Software RAID5 array. One of the drives has a problem and the metadata of the other drives is probably corrupt.
When I try to add a missing disk I get the following error with either Auto Detect or Advanced Detect.

"Failed to reconstruct RAIDs structure. Unsupported RAID type or invalid data.

If I add the real 4th drive instead, it begins detecting with it and then drops it a minute or two later.

Is there anything I can do about the error?
Thanks

Hello Hobble

Read my posts on this forum regarding RAIDs and precautions to take.

Next

are you sure the other 3 drives were all up, running and still employed by the RAID?

Hi Corsari,

Thanks, I had actually found most of your old posts before posting. This particular event was caused by a metadata update that went wrong. md 0.9 -> 1.0 -> 1.2. All the drives are physically fine. R-Studio will start to search for the RAID parameters with the 4 drives no problem, although it drops the one that has a gpt very quickly. I really wanted to try and re-construct with the 3 remaining drives and a "missing disk" I couldn't find any reference to the error I am getting in your posts (maybe I didn't look hard enough).

I can't click on auto-detect without getting the error if a missing disk is in the mix.
Could you please point me in the direction of your post that would help?
Thanks
HobbledGrubs

actually we don't recover linux RAIDs with mdadm commands.

So from here on it's your job

Though, unless you had originally not created the RAID with some fancy custom parameter, you'd enter in R-Studio the default mdadm RAID parameters and check for consistency of some common files types you had in your storage.

Which was it the filesystem? EXT, UFS, XFS, ZFS etc

when you say "we" do you mean R-Studio or your company?

The system was ext4. There was nothing special about it at all except that one drive had been replaced by a much larger drive using a GPT.

I mean @ our company (it is Italy based)

What do you mean with replaced by a larger one using a GPT? Anyway it is not relevant, if you confirm that this replacement has been done previously than it raised the current issue and that the system had worked fine for some time with the 3 original drives + the replaced larger one.

Yep the system was working fine beforehand.
I am trying to calculate a better sector offset to get detection to work. 30hrs of detection is turning up nothing but the sector alignments are getting further and further apart. The sector alignment should be synchronised.